260 likes | 474 Views
Kien A. Hua Data Systems Lab Division of Computer Science University of Central Florida. Data Systems Lab. Data Management. Data Privacy & Security. Data Systems Lab. Data Understanding. Data Communications. Traditional Internet Users - Human. This is changing.
E N D
Kien A. Hua Data Systems Lab Division of Computer Science University of Central Florida
Data Systems Lab Data Management Data Privacy & Security Data Systems Lab Data Understanding Data Communications Data Systems Lab, Division of Computer Science
Traditional Internet Users - Human This is changing Data Systems Lab, Division of Computer Science
“Things” are becoming smart !! Thanks to all kinds of sensors Data Systems Lab, Division of Computer Science
Internet of Things Emerging new users of Internet - “things” ! There will be more of them than us - a lot more !! Je parle digital My feeling is analog Data Systems Lab, Division of Computer Science
IoT Technology • Advances in hardware outpaces those in software • What is a good platform for Internet-scale sensor computing ? Data Systems Lab, Division of Computer Science
Internet of Camera Things • The city of London alone has more than 10,000 cameras • Essentially every adult carries a couple of cameras all the time, i.e., smart phone What is a good platform for Internet-scale live video computing ?? Huge market potential ! Data Systems Lab, Division of Computer Science
Live Video Computing (LVC) If we treat camera as a special class of storage device, old tricks still work. Conventional Computing Live Video Computing Process live video feeds Process data captured on disks Data Systems Lab, Division of Computer Science
Database Approach to LVC Ad hoc Queries Developers focus on application logic Application 7 Application 1 Application 2 Application 3 Application 4 Application 5 Application 6 Application 8 Query Interface Data access supported by Live VDBMS through event-based search Live VDBMS Live video feeds from a new class of storage devices Data Systems Lab, Division of Computer Science
Database Approach to LVC A Web service is a software function provided at a network address on the Web Ad hoc Queries Application 7 Application 1 Application 2 Application 3 Application 4 Application 5 Application 6 Application 8 Query Interface Put all of these in a Web service Live VDBMS Data Systems Lab, Division of Computer Science
An Internet Platform for LVC LVC Service LVC Service LVC Service LVC Service Multimedia meets Big Data The Internet-scale continuous data sets are too large and complex for traditional database tools Data Systems Lab, Division of Computer Science
Privacy Certification An executableprivacy specification language allows for a formal way to design, verify, test, and deploy privacy policies Having a standardized privacy specification language is an important step toward privacy certification Data Systems Lab, Division of Computer Science
A High-Performance “Green” Internet • Observation: • 90% of Internet traffic is video • 10% of videos account for 90% of video accessed at YouTube → a lot of “redundant” transmission • Internet accounts for 2% of worldwide energy consumption • Opportunity: • Redundancy control conserves Internet bandwidth for emerging applications such as IoT, and saves significant energy Data Systems Lab, Division of Computer Science
Video-on-Demand (VoD) Challenge Multicast: Wait for multicast time. This is not VoD This is what we want: Do not need to wait; but how ? Data Systems Lab, Division of Computer Science
Video Streaming Tree Source connected to destinations as in conventional routing Smart router Data Systems Lab, Division of Computer Science
Video Streaming Tree Smart router reuses data from an older stream for a newer stream - controlling redundancy ! Smart router Merge Data Systems Lab, Division of Computer Science
Video Streaming Tree Merging taking place independently throughout the network incrementally constructs a video streaming tree Data Systems Lab, Division of Computer Science
Video Streaming Tree Merging taking place independently throughout the network incrementally constructs a video streaming tree Data Systems Lab, Division of Computer Science
Video Streaming Tree Merging taking place independently throughout the network incrementally constructs a video streaming tree Data Systems Lab, Division of Computer Science
Video Streaming Tree Controlling redundancy prevents bottlenecks and reduces network traffic Bottleneck Without video streaming tree More traffic Data Systems Lab, Division of Computer Science
Streaming Tree Multicast Tree Multicast: Wait for multicast time. Limited Application Streaming Tree: Video on demand, many more applications Data Systems Lab, Division of Computer Science
Deployment Replace the Internet with the smart routers tomorrow Not going to happen ! Smart Routers
Smart Overlay Network • Smart overlay network consists of smart routers capable of merging redundant streams • The underlying Internet is abstracted and presented as streaming-tree service to video applications Logical link Smart overlay Smart router Internet Logical link Data Systems Lab, Division of Computer Science
Incremental Deployment Smart routers can be gradually added as the old routers are deprovisioned from Internet Physical network Traditional router Data Systems Lab, Division of Computer Science
Prototyping & Experimentation • NS3 simulation done • Prototype Ver 1.0 will be ready soon • We will set up software routers in North America, Asia, and Europe by mid 2014 • We hope to find funding to speed up this effort Business Opportunities • Forming new business unit to provide network service to “video-on-demand” companies • Selling smart routers to network service providers
Database Courses at UCF • COP4710: Fundamental of Database Systems • COP5711: Parallel and Distributed Database Systems • COP6730: Transaction Processing Systems • COP6731: Advanced Database Systems Databases